Haug, Constantin (2019)
Numerical modelling of production-induced stress changes and seismicity in fault-compartmentalized reservoirs of the North German Basin.
Technische Universität Darmstadt
Dissertation, Erstveröffentlichung
Kurzbeschreibung (Abstract)
Natural gas production from some Rotliegend reservoirs in the Netherlands and North Germany has triggered seismic events after the affected sites had undergone a significant decrease in pore pressure. Recent seismic events in North Germany have caused negligible structural damage but major public concern in a region of low historic seismicity. In a first order approach, the pore pressure depletion of deep reservoirs is accompanied by stress changes that superimpose on pre-production and far-field stresses and potentially question the strength of pre-existing fault planes. Basic analytic models can explain how first, the production-induced differential increase of effective vertical and horizontal stress components within the reservoir, and second the compensation of reservoir contraction by compatibility stresses in the reservoir surroundings, can lead to the reactivation of pre-existing faults during depletion. These analytic approaches fail however to account for the mechanical stratigraphy and structural complexities occurring in reality. In this study, general characteristics of Rotliegend gas fields were depicted in 2D Finite element models in order to improve the quantitative understanding of production-induced seismicity and identify geological parameters that may favour the occurrence of seismicity in North Germany. The tendency towards normal faulting was investigated for a parameter space of reservoir depth, reservoir thickness, mechanical and hydraulic properties and compartment geometries. Distinct features of the local geology such as a varying thickness of viscoelastic salt in the overburden and offset reservoir compartments were addressed by three basic model settings whereby a graben setting served as a reference scenario. The consolidation procedure of the FE- multiphysics tool ABAQUS was used to solve the fully coupled poroelastic equations, requiring the establishment of a modelling strategy first. Herein the Biot-coefficient, that commonly assumes values inferior to one in solid rocks, constituted a crucial model parameter. In the numerical procedure, the evaluation of failure and output of Terzaghi effective stress ignored the Biot-coefficient at first glance. It is however shown that the coefficient scales with total stresses and governs the magnitudes of both effective stress concepts, the Biot-Willis effective stress, reflecting the direct poroelastic stress-strain relation, and Terzaghi effective stress determining the yield of porous rock formations under large confining pressures. A second crucial feature, the deterministic model fault was depicted by contact surfaces that capture the discontinuous character of real faults and allow for relative displacements. In the poroelastic modelling, sealing faults invoke continuity constraints that, along with an inconvenient definition of failure stress, lead to the introduction of the fault-loading parameter SSR, used for the comparative evaluation of fault-loading in different model variations. Identified critical fault-loading patterns were in a second modelling campaign allowed to dissipate by means of dynamic slip, whereby the implementation of slip-rate weakening friction and the consideration of dynamic stresses in ABAQUS explicit captured essential characteristics of dynamic rupture. Several factors that may favour production-induced seismicity in North Germany were identified in the modelling. First, individual parameters such as large reservoir thickness, large Biot-coefficients of the reservoir, shallow reservoir position, a large stiffness contrast between caprock and reservoir and in particular the effect of viscoelastic evaporites in the overburden, exerting a non-homogeneous load and inhibiting stress redistribution, are factors favouring the reactivation of the graben-bounding fault. Second, intra-field compartmentalizing faults are found a more likely location for seismic events than the graben boundary. Particularly production scenarios that deplete either both of two offset compartments, or the footwall compartment, exert a strong loading on the fault. Furthermore, partly-juxtaposed reservoir compartments constitute a preferential geometric setting for fault reactivation. A favoured reactivation of steeply dipping faults (> 60°) in all models reflects a dominant contribution of compaction-strains to fault-loading and contrasts with tectonic concepts. The fault reactivation potential on the graben boundary is found to be highest at the upper reservoir level and a narrow interval above it. For the intra-field setting, rupture initiates at the lower level of the fault section bounding the footwall compartment. Rupture generally propagates either downwards or simultaneously up– and downwards. Thereby, the interplay of the overall stress state and the initial acceleration can favour or inhibit slip propagation but the isotropic stress state within the salt and the tendency for reverse faulting below the reservoir pose ultimate geologic barriers to the propagation of rupture. The final rupture pattern depends predominantly on the geological setting and its local stress state, whereas the effect of the simulated depletion scenarios is hardly distinguishable. Bulk formation strain is shown to be ambiguously expressed by two different effective stress concepts, moreover, the modelling results reveal a strong sensitivity on the representation of the fault. These findings of the study highlight specific challenges of the modelling of reservoir geomechanical problems. Reviewing the model results, a single dominant parameter for seismic hazard cannot be derived as fault criticality is predominantly governed by a complex interplay of initial stresses, pore pressure and fault strength. The investigated parameters revealed the potential to enhance or mitigate critical tendencies in an interfering manner. In consequence, more detailed prediction models require more detailed structural and field data. Nevertheless, the FE models enhance the understanding of processes and influential parameters for induced seismicity and are able to investigate scenarios and define limiting cases in the planning and operation of fields. The presented two-step approach, encompassing the prioritized identification of critically stressed faults and the separate simulation of rupture can be extended to case investigations addressing for example problems of caprock integrity.
Typ des Eintrags: | Dissertation | ||||
---|---|---|---|---|---|
Erschienen: | 2019 | ||||
Autor(en): | Haug, Constantin | ||||
Art des Eintrags: | Erstveröffentlichung | ||||
Titel: | Numerical modelling of production-induced stress changes and seismicity in fault-compartmentalized reservoirs of the North German Basin | ||||
Sprache: | Englisch | ||||
Referenten: | Henk, Prof. Dr. Andreas ; Alber, Prof. Dr. Michael | ||||
Publikationsjahr: | 11 Februar 2019 | ||||
Ort: | Darmstadt | ||||
Datum der mündlichen Prüfung: | 19 Dezember 2018 | ||||
URL / URN: | https://tuprints.ulb.tu-darmstadt.de/8476 | ||||
Kurzbeschreibung (Abstract): | Natural gas production from some Rotliegend reservoirs in the Netherlands and North Germany has triggered seismic events after the affected sites had undergone a significant decrease in pore pressure. Recent seismic events in North Germany have caused negligible structural damage but major public concern in a region of low historic seismicity. In a first order approach, the pore pressure depletion of deep reservoirs is accompanied by stress changes that superimpose on pre-production and far-field stresses and potentially question the strength of pre-existing fault planes. Basic analytic models can explain how first, the production-induced differential increase of effective vertical and horizontal stress components within the reservoir, and second the compensation of reservoir contraction by compatibility stresses in the reservoir surroundings, can lead to the reactivation of pre-existing faults during depletion. These analytic approaches fail however to account for the mechanical stratigraphy and structural complexities occurring in reality. In this study, general characteristics of Rotliegend gas fields were depicted in 2D Finite element models in order to improve the quantitative understanding of production-induced seismicity and identify geological parameters that may favour the occurrence of seismicity in North Germany. The tendency towards normal faulting was investigated for a parameter space of reservoir depth, reservoir thickness, mechanical and hydraulic properties and compartment geometries. Distinct features of the local geology such as a varying thickness of viscoelastic salt in the overburden and offset reservoir compartments were addressed by three basic model settings whereby a graben setting served as a reference scenario. The consolidation procedure of the FE- multiphysics tool ABAQUS was used to solve the fully coupled poroelastic equations, requiring the establishment of a modelling strategy first. Herein the Biot-coefficient, that commonly assumes values inferior to one in solid rocks, constituted a crucial model parameter. In the numerical procedure, the evaluation of failure and output of Terzaghi effective stress ignored the Biot-coefficient at first glance. It is however shown that the coefficient scales with total stresses and governs the magnitudes of both effective stress concepts, the Biot-Willis effective stress, reflecting the direct poroelastic stress-strain relation, and Terzaghi effective stress determining the yield of porous rock formations under large confining pressures. A second crucial feature, the deterministic model fault was depicted by contact surfaces that capture the discontinuous character of real faults and allow for relative displacements. In the poroelastic modelling, sealing faults invoke continuity constraints that, along with an inconvenient definition of failure stress, lead to the introduction of the fault-loading parameter SSR, used for the comparative evaluation of fault-loading in different model variations. Identified critical fault-loading patterns were in a second modelling campaign allowed to dissipate by means of dynamic slip, whereby the implementation of slip-rate weakening friction and the consideration of dynamic stresses in ABAQUS explicit captured essential characteristics of dynamic rupture. Several factors that may favour production-induced seismicity in North Germany were identified in the modelling. First, individual parameters such as large reservoir thickness, large Biot-coefficients of the reservoir, shallow reservoir position, a large stiffness contrast between caprock and reservoir and in particular the effect of viscoelastic evaporites in the overburden, exerting a non-homogeneous load and inhibiting stress redistribution, are factors favouring the reactivation of the graben-bounding fault. Second, intra-field compartmentalizing faults are found a more likely location for seismic events than the graben boundary. Particularly production scenarios that deplete either both of two offset compartments, or the footwall compartment, exert a strong loading on the fault. Furthermore, partly-juxtaposed reservoir compartments constitute a preferential geometric setting for fault reactivation. A favoured reactivation of steeply dipping faults (> 60°) in all models reflects a dominant contribution of compaction-strains to fault-loading and contrasts with tectonic concepts. The fault reactivation potential on the graben boundary is found to be highest at the upper reservoir level and a narrow interval above it. For the intra-field setting, rupture initiates at the lower level of the fault section bounding the footwall compartment. Rupture generally propagates either downwards or simultaneously up– and downwards. Thereby, the interplay of the overall stress state and the initial acceleration can favour or inhibit slip propagation but the isotropic stress state within the salt and the tendency for reverse faulting below the reservoir pose ultimate geologic barriers to the propagation of rupture. The final rupture pattern depends predominantly on the geological setting and its local stress state, whereas the effect of the simulated depletion scenarios is hardly distinguishable. Bulk formation strain is shown to be ambiguously expressed by two different effective stress concepts, moreover, the modelling results reveal a strong sensitivity on the representation of the fault. These findings of the study highlight specific challenges of the modelling of reservoir geomechanical problems. Reviewing the model results, a single dominant parameter for seismic hazard cannot be derived as fault criticality is predominantly governed by a complex interplay of initial stresses, pore pressure and fault strength. The investigated parameters revealed the potential to enhance or mitigate critical tendencies in an interfering manner. In consequence, more detailed prediction models require more detailed structural and field data. Nevertheless, the FE models enhance the understanding of processes and influential parameters for induced seismicity and are able to investigate scenarios and define limiting cases in the planning and operation of fields. The presented two-step approach, encompassing the prioritized identification of critically stressed faults and the separate simulation of rupture can be extended to case investigations addressing for example problems of caprock integrity. |
||||
Alternatives oder übersetztes Abstract: |
|
||||
URN: | urn:nbn:de:tuda-tuprints-84760 | ||||
Sachgruppe der Dewey Dezimalklassifikatin (DDC): | 500 Naturwissenschaften und Mathematik > 550 Geowissenschaften | ||||
Fachbereich(e)/-gebiet(e): | 11 Fachbereich Material- und Geowissenschaften 11 Fachbereich Material- und Geowissenschaften > Geowissenschaften 11 Fachbereich Material- und Geowissenschaften > Geowissenschaften > Fachgebiet Ingenieurgeologie |
||||
Hinterlegungsdatum: | 10 Mär 2019 20:55 | ||||
Letzte Änderung: | 10 Mär 2019 20:55 | ||||
PPN: | |||||
Referenten: | Henk, Prof. Dr. Andreas ; Alber, Prof. Dr. Michael | ||||
Datum der mündlichen Prüfung / Verteidigung / mdl. Prüfung: | 19 Dezember 2018 | ||||
Export: | |||||
Suche nach Titel in: | TUfind oder in Google |
Frage zum Eintrag |
Optionen (nur für Redakteure)
Redaktionelle Details anzeigen |